[Framework-Team] Re: plone.app.locking

Martin Aspeli optilude at gmx.net
Sun Dec 10 17:53:25 UTC 2006


Hi Raphael,


>> To get the UI to work, I want to provide a couple of viewlets that (a) 
>> trigger the auto-lock when base_edit is found (b) remove the auto-lock 
>> when the base_edit form is unloaded (probably in JS) and
> 
> That's exactly what Jeff and I did and why it's broken ATM
> (I think at least; doing AJAX on top of Zope 2 versus Zope 3's
> templating engine seem to be incompatible). :-(
> (Please, please prove me wrong ...)

I can't see any viewlet related code anywhere. Note that viewlets work 
fine (we use them in other places) so I don't know why this would be a 
problem.

I'm not quite sure what the KSS dependency may be. It would be nice if 
stealing a lock didn't mean reloading the page, and there may be other 
things I'm not thinking about, but I don't think there's anything that 
couldn't be fixed up later. I'm probably missing something, though.

Also, the LockManager product was not in the bundle (though it's 
referenced in the code and can be found in the Collective). I don't 
understand where it's needed, though.

>> (c) provides the lock status/lock stealing form.
>>
> 
> Basically this is all there. Seems to me that we are only talking
> about some code re-arragnement.

That is precisely what this is, but the arrangement of the code was the 
main criticism the first time around, and we need to make things move.

> iterate also uses the WebDAV locks but the suggestion has been that
> it should be harder if not impossible to steal a lock issued by iterate.
> For that, we might need something like lock levels (or whatever)
> in order to be able to distinguish locks.

Okay. I think any such code should live in plone.app.locking (or 
similar), i.e. a common dependency, rather than AT or iterate, though. 
It would really just be a policy decision in the UI for stealing locks 
in the end.

>> I would *like* to just commit plone.app.locking and merge the 
>> remaining support into AT and Plone without re-branching and 
>> re-bundling, to save time. I can do so if people ahve reservations, 
>> though.
>>
> +1 from me. (just do it ...)

Thanks :)

Martin





More information about the Framework-Team mailing list